... I think the logic outlined in the Royal Society paper would still apply.
I.E the motion of the balances in perfect anti-phase should be the least dampened, because the balances movement at the point at which they are close (in the middle of the movement) would always be in the same direction. Conversely, if they are out of phase, there will be times (if they are only a little out of phase) in which the movement is opposite, and there should be friction then because the air turbulence is in the same direction as the balances movement and hence, in these time points opposite. That means that anti-phase synchrony will remain the only stable mode because all the others are dampened out. Does that make sense?
If this is correct, here is the next thing I am wondering about: the Journe beats at 3Hz (21600 vph). So every 1/3 of a second the balance wheel will be in the same position. I think this would mean that if there is an external disturbance that causes more than 1/3 of a second deviation I do not see how the balance wheel would be moved, by coupled oscillation, back to 0 sec deviation - I think it should stay at 1/3. Same for 2/3, 1 sec, etc.
However, I am watching the watch really closely (by a massive amount of willpower I now sometimes manage not to stare at it when in conversation, or driving fast. Getting better) and I do not see that happening. The watch is either in Resonance or the deviation is suddenly more than 1 sec, and doesn't go back. Any other experiences?
Best
Andreas